April 3: The Next MineThatData Happy Hour!

Let's go with 8:00pm EDT / 5:00pm PDT, 55 minutes. Send me an email message (kevinh@minethatdata.com) if you want to join us, and I'll send you a link. 

Where Should I Implement An Action Stream?

Here's a customer who just purchased for the fifth time ... via models, the customer is expected to become less responsive with each month that passes.

In my projects, I leverage "Action Segment Levels" to dictate when Action Streams are implemented. Among twelve-month buyers?

• 75% Annual Rebuy Rate = Elite Customers.
• 60% Annual Rebuy Rate = Loyal Customers.
• 40% Annual Rebuy Rate = Quality Customers.
• 20% Annual Rebuy Rate = Average Customers.
• 1%+ Annual Rebuy Rate = Struggling Customers.

Among lapsed buyers?

• 15% = Lapsed Spend Money
• 5% = Lapsed Experiment
• < 5% = Lapsed Save Money

Looking at the graph above, there are several places where Action Streams make sense.

• Month 4 when the customer is about to drop out of Loyal status.
• Month 8 when the customer is about to drop out of Quality status.
• Month 11 when the customer becomes responsive due to an Anniversary event.
• Month 17 when the customer is about to drop out of Lapsed Spend Money status.
• Month 23 when the customer becomes responsive due to an Anniversary event.
• Month 33 when the customer is about to drop out of Lapsed Experiment status.

In those months, you work outside of your normal email/sms program, kicking off automated campaigns designed to keep the customer at the higher level of responsiveness. Most email service providers have these capabilities, they're just waiting for you to use them! Hire me to perform the research, we'll have your ESP implement my equations and we're off and running!

Low Lifetime Value

In maybe 80%(ish) of my projects, there are a handful of high-selling items that bring in a disproportionate number of new customers.

Those items, of course, frequently bring in new customers that have low lifetime value.

I once worked with a Chief Marketing Officer who refused those customers ... he wouldn't pay for them. "Let somebody else have them, I only want high-value customers" he said to me. His business, of course, struggled to grow.

I'm perfectly fine with you paying less for those customers. Have at it.

I'm even better with you setting up an Action Stream for your new customers. Get busy! Identify the items on a second order that launch some of the low-value customers to average-value status, and market those items in your Action Streams.

It's not hard to do.

It makes common sense.

So go do it!

Discussing Lands' End with ChatGPT

Y'all realize, of course, that your utter dependence on Facebook/Google for new customers puts the future health of your business at serious risk, right?

Here's a discussion I had with ChatGPT.

Kevin:  Lands' End seems to rely heavily on discounts and promotions as part of its core business strategy for several reasons:

1. Perceived Value Strategy

Constant discounts create a sense of getting a deal, which encourages customers to buy more. If customers feel like they’re getting high-quality products at a reduced price, they’re more likely to make a purchase — even if the "original" price was inflated to allow for the discount.

2. Customer Conditioning

Over time, Lands' End has trained its customer base to expect discounts. If you know there’s always a sale around the corner, you might hesitate to buy at full price — but you'll jump at a 40% off email. This creates a shopping pattern where customers are almost waiting for the next sale to make a purchase.

3. Competitive Market Pressure

The apparel industry is highly competitive, especially in the mid-tier, classic, and outdoor clothing markets where Lands' End operates. Competitors like L.L.Bean, Eddie Bauer, and even department stores frequently run sales. Lands' End likely feels the need to match or exceed these offers to stay relevant.

4. Inventory Management

Frequent sales help clear out seasonal and excess inventory. Lands' End sells a lot of seasonal clothing (coats, swimsuits, etc.), so offering regular discounts helps prevent stock from piling up and reduces storage costs.

5. Catalog and Email Marketing Dependence

Lands' End has always leaned heavily on direct mail and email marketing. Regular discounts give them a reason to send those emails and catalogs ("Get 30% Off Now!") and drive immediate sales spikes.

6. Margin Cushioning

Lands' End likely sets its initial prices higher, knowing that a large percentage of sales will happen at a discount. That way, even after applying a 30%–50% off promo, they can still maintain a healthy profit margin.

🧥 Bottom Line: It’s a mix of customer psychology, competitive pressure, and strategic pricing. If Lands' End suddenly stopped offering promotions, many customers might stop buying altogether because they’ve been conditioned to expect a discount.

Kevin:  Does Lands' End sell outerwear that would fit me?

That depends — what’s your size and fit preference? Lands' End is known for offering a pretty wide range of sizes, including:

• Regular, Petite, Tall, and Plus for women
• Regular, Tall, and Big & Tall for men
• Some outerwear even comes in extended sizes up to 3X (for women) and 4X (for men)

They’re also known for more relaxed, classic fits rather than slim or trendy cuts. If you prefer a roomy, comfortable fit, you’ll probably have good luck. Plus, they tend to use quality materials (like down, PrimaLoft, and waterproof fabrics) at reasonable prices, especially when you catch one of their sales.

What kind of outerwear are you thinking — a winter coat, rain jacket, or something lighter? 👀

Kevin:  I need a rain jacket, mens, size = 3x.

Got it! Lands' End does carry men's rain jackets in 3X, and their outerwear tends to be pretty size-inclusive. Their Squall® jackets are a popular choice — waterproof, windproof, and lined for extra warmth. They also have some lighter rain shells if you’re looking for something less insulated.

Want me to check what’s available right now or help narrow down the options? 🌧️😊

Kevin: Yes, and please recommend cheaper options from competing brands (links and images were provided, FYI).

Certainly! Here are some men's rain jacket options in size 3X, including selections from Lands' End and more affordable alternatives from other brands:

Lands' End Men's Big Waterproof Rain Jacket
This jacket features a wind-and-waterproof 2.5-layer shell with a breathable, moisture-wicking interior membrane. It includes an adjustable hood, front and chest zipper pockets, and underarm zippers for ventilation. Available in size 3X.

Portwest US440 Classic Waterproof Rain Jacket
An affordable option, this jacket is waterproof with sealed seams and a pack-away hood. It's available in olive color, size 3X.

Frogg Toggs Men's Rain Suit
This rain suit includes both a jacket and pants, offering waterproof protection. Frogg Toggs is known for lightweight and breathable rainwear. Sizes up to 3X are available. 

Columbia Men's Rain Jackets
Columbia offers a range of waterproof and breathable rain jackets. Some styles are available in extended sizes, including 3X.

REI Co-op Men's Rain Jackets
REI carries various men's rain jackets from brands like Marmot and The North Face. Some options are available in size 3X.

When selecting a rain jacket, consider factors like waterproofing, breathability, and additional features such as adjustable hoods and ventilation options. Lands' End offers quality options, and brands like Portwest and Frogg Toggs provide more budget-friendly alternatives without compromising on essential features.

Now, I'm asking questions like a moron, not like a human being. But please pay attention to the judgments that ChatGPT makes about Lands' End and other companies (calling Portwest budget-friendly).

Imagine the content that ChatGPT had to "steal" to train itself to make judgments about who Lands' End is, who they compete with, and the give-and-take between catalogs / email / discounts?

From a commerce standpoint, Google and Facebook are "dead men walking" ... via their current business model. If you are having conversations with AI and AI is steering you where it wants via judgements made by equations that are easy to create but hard to understand, you won't sit in front of a laptop typing questions on Google then visiting landsend.com, will you?

That personal relationship you have with your customers ... either via the community you created, or via your own email/sms marketing program ... are about to become critically important. Ultimately, your email/sms program is going to be replaced by something interactive, but until then that personal relationship is going to be so darn important.

Also ... what do you do when AI makes judgments about your brand that you don't like?

Gonna be a wild ride.

P.S.: This is the point where some of the hounds come after me ... "Kevin, Meta isn't stealing content, they're using readily available information to train AI to be smarter." Well, I searched LibGen to see if Meta used my information, and yes, they used several of my booklets to train their AI. Which, of course, is stealing. And yes, there's nothing I can do about it so complaining is meaningless. I have to adapt and move on. ChatGPT tells me that it makes a distinction between creating new ideas and synthesizing existing ideas. Oh. Great. Also faulty logic.

#16 Seed Upsets the #1 Seed

It's NCAA tournament time, women and men. Sixty-eight teams are whittled down to just one, one heartbreaking moment at a time.

You have a veritable plethora of new products. What stops you from creating a tournament bracket of sixteen new items, and then for fifteen days you pit those items against each other via email marketing ... rank order them based on predicted volume or just assign random seeds, whatever you want. Schedule the matchups, one per day, on March 24 / 25 / 26 / 27 / 28 / 29 / 30 / 31.

• Day 1 = #16 vs. #1.
• Day 2 = #9 vs. #8.
• Day 3 = #13 vs. #4.
• Day 4 = #12 vs. #5.
• Day 5 = #15 vs. #2.
• Day 6 = #10 vs. #7.
• Day 7 = #14 vs. #3.
• Day 8 = #11 vs. #6.

There's your first round. The items that generate the most sales for that day advance to the next round. Schedule these matchups on April 1 / 2 / 3 / 4

• Day 9 = Day 1 Winner vs. Day 2 Winner.
• Day 10 = Day 3 Winner vs. Day 4 Winner.
• Day 11 = Day 5 Winner vs. Day 6 Winner.
• Day 12 = Day 7 Winner vs. Day 8 Winner.

After Day 12 you have whittled the field down to the Final Four! It's time for the semi-finals. Schedule these matchups on April 5 and April 6.

• Day 13 = Day 9 Winner vs. Day 10 Winner.
• Day 14 = Day 11 Winner vs. Day 12 Winner.

Now we're down to the Championship! Schedule it on April 7, the day of the actual Men's National Championship Game.

• Day 15 = Day 13 Winner vs. Day 14 Winner.

The whole thing will last you two weeks. During those two weeks, you feature sixteen new items, with the best sellers being featured three or four times. Maybe the #16 seed will upset the #1 seed, who knows? Your customers will decide.

What is the downside of doing this?

P.S.: I'm told one of the downsides is "telling the competition how many units you are selling". My goodness. Nonsense. Have you ever performed a search on Amazon? Try this one for the Sennheiser HD600. They tell the world they sold at least 400 of 'em in the past month. They also sold more than a thousand of the lower cost (but still good sounding) HD560S according to the search. The company pounding you into submission gladly shares how many units they sell.

P.P.S.:  Somebody will tell you not to do this ... "people will just buy items and return them to game the outcome." Yeah, there probably are five customers who will do this. Now you have a data attribute to append to your "customer data lake" to make sure you never spend money marketing to those customers again.

She Said She Liked AI Because "It Worked"

Her company hired a vendor, and the vendor did a really nice job of building a model to determine which customers they should target.

I asked her, "What are the attributes of the customers that the model preferred?"

She said, "I don't know and I don't care, it worked."

With every advance in technology, we gain something and we lose something.

In this case, her company gained profit.

Of course, she lost knowledge.

There is going to be an opportunity over the next five years to be the person who "knows things". As we embark on a mission to transfer knowledge from employees to computers, the employee who "knows things" will be able to solve problems. For instance, what happens when AI tells you that you should reactivate a specific cohort of customers, and eighteen months later those customers have no future value? Do you blame AI? Or does the employee who "knows things" protect the company?

Be the person who "knows things".

An Offer For Blog Readers: Pricing Analysis

I've had several inquiries in the past two weeks about the impact of prices on customer behavior.

Specifically, you are asking what happens to customer behavior if prices increase by 10% to 20%, through no fault of your own.

We have a lot of data from 2021-2022 to guide us in this endeavor. Here's the process I went through three years ago to address the issue.

1. Let's say you have 15 merchandise categories. When you send me five years of purchase history, I have four years of year-over-year pricing comparisons times fifteen merchandise categories, for sixty (60) data points. That's enough for me to do some damage.
2. I'll compare year-over-year impact of prices (say, +10%) on rebuy rates (say, -5%), annual orders per buyer (maybe flat), annual items per order (say -2%) ... which leads us to what I call demand per inventory (-7%). If prices increase by 10% and customer productivity decreases by 7%, you know that the top-line won't be hurt, but you'll have 5% fewer customers repurchasing, which means you'll have a contracting business.
3. I'll perform the same analyses on new/reactivated buyers. Pricing issues typically hurt new/reactivated customers more than they hurt existing buyers.
4. Given what I observe, I'll forecast the next few years as a consequence. You'll have a good/directional idea whether price increases will hurt your business.

The core part of the analysis that informs the work I perform is in orange-ish cells below.

Let's try something ... I'll do this for $4,990 as a test, prepaid. This is likely going to be way too little to charge for something like this and I'm going to feel like I ripped myself off. It's not a typical $15,000 pricing project, of course, but for nearly free you get to learn the likely impact pricing increases would have in your business, should they come to fruition.

Contact me now (kevinh@minethatdata.com) ... offer ends on Friday.